From: Daniel Vetter Subject: Re: [PATCH 1/2] drm/i915: Add a partial instruction shootdown workaround on Broadwell. Date: Wed, 5 Mar 2014 15:48:20 +0100 Message-ID: <20140305144820.GA17001@phenom.ffwll.local> References: <1393487971-739-1-git-send-email-kenneth@whitecape.org> <20140227084324.GG3852@intel.com> <20140228000504.GA16719@bwidawsk.net> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: Received: from mail-ea0-f177.google.com (mail-ea0-f177.google.com [209.85.215.177]) by gabe.freedesktop.org (Postfix) with ESMTP id C92FAFA748 for ; Wed, 5 Mar 2014 06:48:24 -0800 (PST) Received: by mail-ea0-f177.google.com with SMTP id h10so1076398eak.22 for ; Wed, 05 Mar 2014 06:48:23 -0800 (PST) Content-Disposition: inline In-Reply-To: <20140228000504.GA16719@bwidawsk.net> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: intel-gfx-bounces@lists.freedesktop.org Errors-To: intel-gfx-bounces@lists.freedesktop.org To: Ben Widawsky Cc: intel-gfx@lists.freedesktop.org List-Id: intel-gfx@lists.freedesktop.org On Thu, Feb 27, 2014 at 04:05:04PM -0800, Ben Widawsky wrote: > On Thu, Feb 27, 2014 at 10:43:24AM +0200, Ville Syrj=E4l=E4 wrote: > > On Wed, Feb 26, 2014 at 11:59:30PM -0800, Kenneth Graunke wrote: > > > I believe this will be necessary on production hardware. > > > = > > > Signed-off-by: Kenneth Graunke > > > --- > > > drivers/gpu/drm/i915/i915_reg.h | 3 +++ > > > drivers/gpu/drm/i915/intel_pm.c | 4 ++++ > > > 2 files changed, 7 insertions(+) > > > = > > > I just realized tonight that my workarounds series never got merged. > > > = > > > After reviewing Ben and Chris's comments, I agree with all of them, so > > > I've dropped the unnecessary patches. I also believe I shouldn't need > > > the pre-production workarounds I sent last time, so I've dropped thos= e. > > > = > > > These two should apply to production hardware, so we probably want th= em. > > > = > > > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i= 915_reg.h > > > index ad75ff7..f36d5e0 100644 > > > --- a/drivers/gpu/drm/i915/i915_reg.h > > > +++ b/drivers/gpu/drm/i915/i915_reg.h > > > @@ -5050,6 +5050,9 @@ > > > #define GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE (1<<10) > > > #define GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE (1<<3) > > > = > > > +#define GEN8_ROW_CHICKEN 0xe4f0 > > > +#define PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE (1<<8) > > > + > > > #define GEN7_ROW_CHICKEN2 0xe4f4 > > > #define GEN7_ROW_CHICKEN2_GT2 0xf4f4 > > > #define DOP_CLOCK_GATING_DISABLE (1<<0) > > > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/i= ntel_pm.c > > > index 4f01b04..df8ad21 100644 > > > --- a/drivers/gpu/drm/i915/intel_pm.c > > > +++ b/drivers/gpu/drm/i915/intel_pm.c > > > @@ -4838,6 +4838,10 @@ static void gen8_init_clock_gating(struct drm_= device *dev) > > > /* FIXME(BDW): Check all the w/a, some might only apply to > > > * pre-production hw. */ > > > = > > > + /* WaDisablePartialInstShootdown */ out of curiousity: Does ROW_CHICKEN survive a gpu reset? We've had fun with this kind of stuff before ... > > = > > +:bdw > > = > > Apart from that: > > Reviewed-by: Ville Syrj=E4l=E4 > > = > = > I think they're both stable material. > And: > Reviewed-by: Ben Widawsky Both merged to dinq. bdw in 3.13 is prelim, so would only need to go into 3.14-fixes. But there's too much bdw fixes floating (iirc 10 patches) that some time to settle everything looks like the right approach. I guess as the platform owner you're then signed up to assemeble a backports branch for 3.14-bdw. -Daniel -- = Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch